ProjectAbstract Cellfateengineering,forexamplethedirecteddifferentiationofpluripotentstemcellsor thedirectconversionamongsomaticcelltypes,holdsgreatpromisetoimprovedisease modeling, drug screening, and to lead to regenerative medicine therapies. I recently developed a novel analytical tool, CellNet, which assesses how well engineered cells approachtheirinvivotargetcelltypesbasedoncellandtissuespecificgeneregulatory networks (GRNs). By applying CellNet to all cell fate engineering studies for which compatible data was available, I discovered several issues that were common to virtuallyallmethodsandtargetlineages.First,Ifoundthattheonlyrobustlyfaithfulfate engineeringwasthatofreprogrammingtopluripotency.Second,IfoundthattheGRNof the starting cell type (e.g. pluripotent stem cells in cases of directed differentiation or often fibroblasts in cases of direct conversion) is often partially retained in engineered cells.Third,IfoundthatGRNsofalternatelineages(i.e.thosenotassociatedwiththe starting cell type or the target lineage) are frequently established in engineered cells. Finally, I found that the complex signaling milieu of the mouse microenvironment to whichengineeredcellsaretransplantedpotentlyrepressesalternate/aberrantlineages and induces the target cell type GRNs in directly converted cells. These observations have revealed several fundamental barriers to faithful cell fate engineering and they define opportunities for progress. My current research program, which I seek to fund through this MIRA opportunity, is to develop novel theoretical and computational methodstodefinecelltypeidentityfromsinglecellRNA-Seq(scRNA-Seq)datawithan emphasisondevelopmentalcelltypesthatemergeduringmesodermdevelopmentand subsequent commitment to chondrocyte fate. As part of this work, we will generate scRNA-Seq data of the developing and adult synovial joint, we will harvest and incorporate publicly available and collaborator-provided scRNA-Seq data of other lineages to make a generally applicable platform for assessing cell type identity at the single cell level of resolution, and we will make the resulting methods, software, and datafreelyavailable.Finally,wewillusethesystemtodeterminetheextenttowhichthe cell types and compositions of directly differentiating mouse ESCs match their in vivo counterparts.